hislaziness

I was expecting something like this but better - https://www.nationalgeographic.com/science/article/trapped-a...

The post is gone...

A more efficient way to infer with lower memory requirements.

From the course page - Classes are in person on Wednesdays from 2:00PM - 4:00PM Eastern Standard Time in E14-526

I do not see any of the AI processors like the Google TPU. Would they be cheaper?

Does it mean DOS/XP are not supported anymore? That would be sad.

Works fine in Asia.

I believe the Title is typically not written by the author. I would guess the Title is optimized for clicks.

Not to digress but we are not able to regulate / control the use of genetic engineering, how can we regulate the current GenAI hype?

Geothermal energy should be enough to sustain life.

Start building real-time data apps in minutes Connect any data source, combine them in real-time and instantly get low-latency Data APIs. All with just a simple configuration!

Papers would be good next step. The concept and the algorithms do not seem very complex. The training and fine tuning is what needs effort and infrastructure.

I really liked this series. https://karpathy.ai/zero-to-hero.html

A bit too late I guess. The only way this will stop is when people realize their expectations were misaligned and things start failing spectacularly.

Where can we find this type of data for other countries? The data here is from the Office of the Australian Information Commissioner (OAIC).

If anyone is interested PAIP is downloadable at https://github.com/norvig/paip-lisp

Fascinating story. Daemon domes from Maxwell's thought experiment.

"For this new analysis, ATLAS physicists revisited its data collected in 2011 at a centre-of-mass energy of 7 TeV (corresponding to 4.6 fb-1, also used in ATLAS’ previous measurement). Researchers employed improved statistical methods and refinements in the treatment of the data, enabling them to reduce the uncertainty of their mass measurement by more than 15%."

It does not have to be "me". My point is we seem to have a different benchmark for Natural Intelligence vs Artificial Intelligence.

For whatever reason we seem to have set a very high expectation from AI as compared to NI (Natural Intelligence). I remember reading "There is a considerable overlap between the intelligence of the smartest bears and the dumbest tourists."

I remember reading this somewhere - "There is a considerable overlap between the intelligence of the smartest bears and the dumbest tourists.". Though I do not think GPT-4 is even close to AGI it can definitely claim to be better at faking it than many intelligent beings can.

Copilot with GPT-4 model, chat and voice, pull requests, command line, and docs to answer questions.

Can someone with access check this for me please. Does bard remember giving a different answer earlier. does it acknowledge that the answer was wrong?

Don't think I understand the argument here. Is it causation vs correlation issue?

Impact of the layoffs?

Where can I read a more scientific treatment of this topic?

Article is from 2021. And the last nuke lost was in 1968. Wondering if all the nukes lost are still viable as a weapon?

Nice collection.

Do we really need a data platform?

Thanks. Debezium looks interesting but it is a bit different than what I am looking for. I am looking for an api approach so that other systems can pull the data without having to worry about is it batch or stream.